Samuel Thibault wrote:
YAMAMOTO, Hiroyuki, le Sun 04 Mar 2007 16:05:01 +0900, a écrit :I suggest that this phenomenon doesn't occur in the gnumach-dbg kernel in fact.Eergl :( Guillem, can you confirm that gnumach and gnumach-dbg are compiled exactly the same way (-O2, notably) except --enable-kdb?
The packages of gnumach and the gnumach-dbg (Version: 2:1.3.99.dfsg.cvs20070211-1) were repackaged with gcc-4.0. In debian/rules file , I fixed it to -O2 instead of the fork of -O0 or -O2, moreover, to make sure, I descrived `--prefix=/usr' in both of gnumach and the gnumach-dbg as the configure option.
Thier patches are not rewritten at all.However, the gnumach-dbg kernel doesn't reproduce the phenomenon on ten tests, although the gnumach kernel does it on test of the first time.
I suggest that we may assert that there is no bug in the gnumach-dbg package as long as we are not so unlucky.
Hiroyuki